The potential of yeast proteins to substitute for traditional fining agents : technological and sensory aspects

(1)

Justus-Liebig-Universität Gießen, Faculty 09, Agricultural Sciences,

Nutritional Sciences and Environmental Management

Hochschule Geisenheim University, Department of Microbiology and

Biochemistry

Doctoral thesis

The potential of yeast proteins to substitute for traditional fining agents –

technological and sensory aspects

submitted

by Bernd Christoph Lochbühler, born in Ulm/Donau, Germany

to Justus-Liebig-Universität Gießen, Faculty 09, Agricultural Sciences, Nutritional Sciences and Environmental Management

(2)

1

Justus-Liebig-Universität Gießen, Faculty 09, Agricultural Sciences,

Nutritional Sciences and Environmental Management

Hochschule Geisenheim University, Department of Microbiology and

Biochemistry

Doctoral thesis

The potential of yeast proteins to substitute for traditional fining agents –

technological and sensory aspects

submitted

by Bernd Christoph Lochbühler, born in Ulm/Donau, Germany

to Justus-Liebig-Universität Gießen, Faculty 09, Agricultural Sciences, Nutritional Sciences and Environmental Management

in partial fulfillment of the requirements of the degree of Dr. agr.

Supervisors:

Prof. Dr. Sylvia Schnell, Justus-Liebig Universität Gießen Prof. Dr. Doris Rauhut, Hochschule Geisenheim University

(3)

2

This work is dedicated to my family

who supported me in any way

(4)

3

Publication of parts of the studies presented in this thesis

Article accepted and published online in November 2014 in “European Food Research and Technology”:

Lochbühler, Bernd; Manteau, Sébastien; Morge, Christophe; Caillet, Marie-Madeleine; Charpentier, Claudine; Schnell, Sylvia; Großmann, Manfred; Rauhut, Doris;

“Yeast protein extracts – an alternative fining agent for red wine”, DOI 10.1007/s00217-014-2373-y, Printed version: April, 2015, vol. 240, issue 4, pp. 689-699

Poster presented at WAC (Wine Active Compounds) – Conference in March 2014 in Beaune, France:

Lochbühler, Bernd; Manteau, Sébastien; Morge, Christophe; Caillet, Marie-Madeleine; Charpentier, Claudine; Großmann, Manfred; Rauhut, Doris;

“Yeast protein extracts as an alternative to classical fining agents for red wines”

Lecture given at the annual meeting of “Forschungsring des Deutschen Weinbaus” (Organization of Research in German Viticulture) in April 2014 in Neustadt/Weinstraße:

“Hefeproteinextrakte als Alternative zu traditionellen Schönungsmitteln bei Rotweinen”(Yeast protein extracts as alternative to traditional fining agents in red wine)

Talk presented at Macrowine Conference 2014 in September 2014 in Stellenbosch, South Africa:

“Yeast proteins – a base of alternative fining agents for red wines”

Oral presentation delivered at OIV (International Organization of Vine and Wine) congress 2010 in November 2014 in Mendoza, Argentina:

(5)

4

List of abbreviations

ATP Adenosine Triphosphate

BCIP 5-Bromo-4-Chloro-3-Indolylphosphate BSA Bovine Serum Albumin

BSE Bovine Spongiform Encephalopathy

DAAD Deutscher Akademischer Austauschdienst (German Academic Exchange Service) DIN Deutsche Industrienorm (German Industry Norm)

DNA Deoxyribonucleic Acid DTT DL-dithiothreitol

EC European Community

EP2 Specific type of molasses

EU European Union

FTIR Fourier Transform Infrared Spectroscopy G Gelatine (only used in figures )

HPLC High Perfrormance Liquid Chromatography HRP Horseradish Peroxidase (only used in figures) IDY Inactive Dry Yeast Preparation

L Lot (only used in graphs)

M Molecular Mass Marker (only used in figures) Meq milli equivalent

NMR Nuclear Magnetic Resonance NTU Nephelometric Turbidity Unit

O.D. Optical Density (used as synonym for absorbance) OIV International Organization of Vine and Wine PAGE Polyacrylamide Gel Electrophoresis

PAS Staining based on oxidation of sugars by Periodic Acid and subsequent Schiff reaction PBS Phosphate buffered saline solution

PMP Polymethyl Pentene PRP Proline Rich Protein PVDF Polyvinylidene Flouride PVP Polyvinyl Pyrrolidone PVPP Polyvinyl Polypyrrolidone RNA Ribonucleic Acid

RPM Rotations per Minute

SB-TI Soybean Trypsin Inhibitor (only used in figures) SDS Sodium Dodecyl Sulphate

(6)

5

List of abbreviations continued

Si Silica sol (only used in figuresand tables) SD Standard deviation (only used in annex) T Tannin (only used in figures and tables) TBS Tris Buffered Saline Solution

Tris Tris (hydroxymethyl) aminomethane

UV Ultra Violet

YEPD Yeast Extract Peptone Dextrose Medium

YP Yeast Product

(7)

6

1 INTRODUCTION ... 10

1.1 Preliminary remarks ... 10

1.2 Objectives and stages of the studies ... 11

1.3 Review of the literature ... 12

1.3.1 Yeast autolysis ... 12

1.3.1.1 General process ... 12

1.3.1.2 Proteolysis... 12

1.3.1.3 Morphological changes in yeast cells during autolysis ... 14

1.3.1.4 Detection of yeast cells being in autolysis ... 15

1.3.1.5 Compounds released during autolysis ... 15

1.3.1.6 Influence of yeast autolysis in wine on its chemical composition and sensory character... 16

1.3.2 Creation of yeast mutants showing inducible autolysis and their use during wine making ... 17

1.3.3 Fining of must and wine ... 17

1.3.3.1 General considerations ... 17

1.3.3.2 Composition of polyphenols of must and wine and their quantification or estimation ... 18

1.3.3.3 Interactions between proteins and polyphenols ... 22

1.3.3.4 Selected studies on fining of must and wine and characterization of protein fining agents ... 25

1.3.4 Use of yeast autolysates or extracts during winemaking and their (potential) sensory effects ... 30

2 THESIS PART I: MUTAGENESIS AND SELECTION OF YEAST STRAINS ... 32

2.1 Material and methods ... 32

2.1.1 Chemical substances ... 32

2.1.2 Media and sterile solutions ... 32

2.1.3 Analysis of EP2 ... 33

2.1.4 pH measurements ... 33

2.1.5 Spectophotometry ... 33

2.1.6 Mutagenesis and first selection of yeast strains ... 33

2.1.7 Yeast strain conservation ... 34

2.1.8 Second selection of yeast strains ... 34

2.1.8.1 Autolysate production ... 34

2.1.8.2 Morphological examination of yeast cells during/after autolysis ... 34

2.1.8.3 Protein analysis in the autolysates ... 34

2.1.8.4 Phenotypic characterization of the strains after heat shock ... 35

(8)

7

2.2 Results ... 37

2.2.1 Mutagenesis and first selection of mutants ... 37

2.2.2 Second selection of yeast strains ... 37

2.2.3 Autolysate of pilot production ... 42

2.3 Discussion ... 44

3 THESIS PART II: FINING OF MUST AND WINES WITH YEAST PROTEIN

EXTRACTS ... 47

3.1.2 General wine analysis ... 47

3.1.3 Spectrophotometric measurements ... 47

3.1.4 Colour measurements of white wines and musts ... 47

3.1.5 Colour measurements of red wines ... 47

3.1.6 Polyphenol measurements of white wines and musts ... 48

3.1.7 Polyphenol measurements of red wines ... 48

3.1.8 Turbidity of musts and wines ... 48

3.1.9 Sensory analysis ... 48

3.1.10 Fining experiments ... 49

3.1.10.1 Cycle 1 ... 50

3.1.10.2 White must of variety Arnsburger (cycle 1) ... 50

3.1.10.3 Cuvee of white wines (cycle 1)... 51

3.1.10.4 Fining of red wines (cycle 1) ... 51

3.1.10.5 Cycle 2 ... 53

3.1.10.6 Experiment with Riesling must (cycle 2) ... 53

3.1.10.7 Experiment with Riesling wine (cycle 2) ... 54

3.1.10.8 Experiment with a cuvee of German red wines (cycle2) ... 54

3.1.10.9 Cycle 3: Fining of red wines: ... 55

3.1.10.10 Cycle 4: Fining of red wines: ... 57

3.1.10.11 Cycle 5: Fining of red wine cuvee ... 58

3.1.10.12 Cycle 6: Fining experiment with commercial YPE 8 ... 59

3.2 Results ... 60

3.2.1 Fining experiments of cycle 1 ... 60

3.2.1.1 Fining of white must of the variety Arnsburger ... 60

3.2.1.2 Fining of a cuvee of white wine ... 60

(9)

8

3.2.1.4 Fining experiments with red wine of Cabernet Sauvignon ... 64

3.2.2 Fining experiment of cycle 2 ... 68

3.2.2.1 Experiment with Riesling must (cycle 2) ... 68

3.2.2.2 Experiment with Riesling wine (cycle 2) ... 68

3.2.2.3 Experiment with cuvee of red wine (cycle 2) ... 69

3.2.3 Fining experiments of cycle 3: fining of a cuvee of red wines ... 69

3.2.4.1 Fining of red wine of Syrah ... 72

3.2.4.2 Fining of red wine of Rondo ... 76

3.2.5.1 Fining of a red wine cuvee at small scale ... 79

3.2.5.2 Fining of red wine cuvee in bigger volumes... 83

3.2.6 Fining experiments with the commercial YPE 8 (experiments of cycle 6) ... 88

3.3 Discussion ... 92 3.3.1 Experiments of cycle 1 ... 92 3.3.2 Experiments of cycle 2 ... 94 3.3.3 Experiments of cycle 3 ... 96 3.3.4 Experiments of cycle 4 ... 97 3.3.5 Experiments of cycle 5 ... 98 3.3.6 Experiments of cycle 6 ... 100

4 THESIS PART III: PARTIAL CHARACTERIZATION OF YEAST PROTEIN

EXTRACTS AND PROTEIN FINING AGENTS ... 102

4.1.2 Quantification of proteins in fining products ... 102

4.1.3 SDS-PAGE of yeast proteins and fining products ... 102

4.1.4 Detection of proteins and glycoproteins on the SDS-PAGE gels ... 103

4.1.5 Detection of glycoproteins on SDS-PAGE gels by Western Blotting ... 103

4.1.6 Detection of mannose and glucose in yeast products ... 103

4.2 Results ... 104

4.2.1 Protein quantification in YPE and fining products ... 104

4.2.2 Distribution of molecular masses of proteins ... 106

4.2.3 Glycoproteins detected in YPE and gelatine ... 110

4.2.3.1 Detection of total glycosylated proteins ... 110

(10)

9

4.2.4 Sugar concentration in selected YPE ... 114

4.3 Discussion ... 115

4.3.1 Quantitative analysis of proteins in YPE and reference protein fining products... 115

4.3.2 Qualitative protein characterization by SDS-PAGE ... 116

4.3.3 Presence of glycoproteins in YPE ... 117

4.3.4 Concentration of sugars in YPE after acid hydrolysis ... 119

4.4 Final conclusions ... 120

5 SUMMARY ... 121

6 ZUSAMMENFASSUNG ... 122

7 ACKNOWLEDGEMENTS ... 123

8 BIBLIOGRAPHICAL REFERENCES ... 125

9 ANNEX ... 136

9.1 Table 1: Morphology of strains after autolysis and protein concentrations in autolysate ... 137

9.2 Table 2: Phenotypical characterization of the strains conserved on plate or deep frozen during after heat shock of 48h at 37°C ... 142

10 DECLARATION OF CONFORMITY OF THE DISSERTATION TO THE RULES OF

THE DOCTORAL EXAMINATION BOARD OF JUSTUS-LIEBIG-UNIVERSITÄT ... 148

(11)

10

1 Introduction

1.1 Preliminary remarks

A study on yeast derivatives used in wine making was the initial project of this thesis.

There was a vast amount of products for oenological treatments based on inactive yeast on the market in 2008 and many were introduced since. Such products are / were for example claiming to provide nutrients to yeast or lactic acid bacteria, to have antioxidant effects on wine or to improve its organoleptic properties. Consequently, it was decided to focus the studies of this thesis on some fields of the vast variety of yeast derived oenological products.

Yeast autolysates were the first products that were examined at the beginning of the studies in 2008 and a contact to Professor Charpentier of “Université de Bourgogne”, in Dijon, France, was established. Professor Charpentier has a big expertise and knowledge on yeast autolysis in wine conditions and the compounds released to wine during that processes. She offered the author the possibility to work some months in her laboratory in 2008 to learn more about the research on yeast autolysis, autolysates etc. Professor Charpentier and her group had started by 2006 a project on protein extracts released by yeast during autolysis.

It was furthermore examined if these extracts could replace traditional products based on animal proteins used for fining of grape must and wine.

A declaration of the use of some fining agents of animal origin, such as out of as milk or egg in wines, towards the consumer was discussed by the legislature of the European Union since 2007 as these proteins can cause allergies in sensitive people. Consequently, alternative sources for proteins for fining, that were not known to cause allergies, were looked for and yeast could be a promising source. Professor Charpentier kindly offered to the author to contribute to that research what was agreed with pleasure.

The aim of studies presented herein was to obtain protein extracts from yeasts in autolysis that were less subject to protein hydrolysis than in the former researches done by other authors (cf. 1.2.).

The declaration of fining of must and wines with products out of egg and milk towards the consumer became finally obligatory in 2012 (regulation (EC) 579/2012). That legal requirement underlined that the research on alternative protein sources for fining agents was a subject meeting current demands of wine professionals.

Furthermore yeast protein extracts are allowed for fining of must and wine in the European Union since 2013 offering an alternative to the wine sector (regulation (EC) No 144/2013).

(12)

11

1.2 Objectives and stages of the studies

The development of protein extracts out of yeasts undergoing autolysis was one of the objectives of the studies of this thesis (cf. section 1.1.).

The other important task was to test if the protein extracts would meet the technological and sensory requirements of a fining agent for grape musts and wines.

Furthermore it would be interesting and necessary to analyze the protein composition of the yeast extracts to control the production process and to gain insight in the composition of promising fining agents out of yeast proteins.

The work of the studies of this thesis was therefore divided in three parts: 1. Mutagenesis and selection of yeast strains

 Objective was to obtain yeasts that autolyse during stress conditions and release high concentrations of proteins

 These proteins should have protein masses above 15 kDa

 Proteins of low degree of hydrolysis were reported to precipitate more completely in wine (Yokotsuka and Singleton, 1987 and 1995) and to have a higher influence on phenolic compounds in wine (Tschiersch et al., 2010), which can be also desired  Work on extraction process to obtain YPE (yeast protein extracts)

2. Fining tests of must and wines with yeast protein extracts.

3. Partial characterization of yeast protein extracts and protein fining agents

Fining of wine always influences its sensorial characteristics (limpidity) and the reaction between proteins of the fining agents with tannins of the wine can also change colour, taste and astringency of wines (Maury et al., 2001).

Physico-chemical and sensorial analyses were used to evaluate the influence of fining with yeast protein extracts on parameters related to wine flavour.

(13)

12

1.3 Review of the literature

1.3.1 Yeast autolysis

1.3.1.1 General process

Autolysis of yeast has been an important subject in biotechnology as well as in food and beverage technology during the last decades. The event has been extensively described in several reviews (Babayan and Bezrukov 1985; Charpentier and Feuillat 1993; Fornairon-Bonnefond et al. 2002). Charpentier and Feuillat (1993) and Fornairon-Bonnefond et al (2002) have dealt with autolysis of yeast in wine and sparkling wine as well as Alexandre and Guilloux-Benatier (2006) in the case of sparkling wine.

The process of autolysis has been well described by Babayan and Bezrukov (1985), a work also cited by the other authors mentioned above. Autolysis is a passive process according to these authors happening after cell death. In a first step, the membranes of the yeast cell have lost their specific permeability and hydrolases are released. These hydrolases degrade intracellular macromolecules after enzyme activation or break down of enzyme inhibitors. Liberation of the hydrolyzed macromolecules out of the cell can occur if cell wall porosity has been increased by partial degradation of its polymers (cf. below). A further transformation of the compounds released by the cell can happen in the surrounding medium by chemical reactions that are partially also catalyzed by enzymes released together with the yeasts’ macromolecules.

Autophagy may also play a role in the process of autolysis in case of yeasts used in food biotechnology as pointed out in the overview of Cebollero and Gonzalez (2007). Autophagy reactions are gene controlled processes taking place in still living cells. Autophagy means self-consumption resulting in transport of compounds and also of organelles of the cytoplasm in autophagosomes, vesicles having a bilayer membrane, to the vacuole. The outer membrane of the autophagosome fuses with the membrane of the vacuole and the inner membrane and the content of the autophagosome are degraded by enzymes. Formation of autophagosomes has been described in yeast cells during the second fermentation of sparkling wine (cf. also below).

1.3.1.2 Proteolysis

Proteolytic activity plays a key role in the release of compounds out of yeast cells during autolysis and will be outlined more in detail, as the release of proteins during yeast autolysis was an important part of the research work of the thesis.

Proteolysis in yeast, especially in Saccharomyces cerevisiae, has been reviewed by different authors (e.g. Achstetter and Wolf 1985, Hilt and Wolf 1992; van den Hazel et al. 1996 and Sorokin et al., 2009). The vacuole of the yeast is an important site of proteases and was supposed to be the place of unspecific broad proteolytic activity playing a role in nitrogen metabolism and stress response (Achstetter and Wolf 1985; van den Hazel et al. 1996). Proteolysis in the vacuole seems to be the basal degradation of long lived proteins and to perform bulk hydrolysis in case of starvation (van den Hazel

(14)

13

et al. 1996). Achstetter and Wolf (1985) described six different enzymes present at the vacuole, among which three were dependent on metal ions, but the authors pointed out that additional protease activity was found without knowing the specific enzymes. Proteinase ysc A and ysc B seem to perform the majority of proteolysis or at least 40 % (van den Hazel et al. 1996) in the vacuole and both enzymes were not reported to be metal dependent (Achstetter and Wolf, 1985; Hilt and Wolf, 1992). Proteinase A with an acidic pH-optimum is located in the vacuole and seems to play an important role in autolysis in wine-like conditions (cf. review of Charpentier and Feuillat 1993). Transport of proteins to be degraded to the vacuole can take place by autophagocytosis in autophagosomes having a double layer of membrane (reviewed by van den Hazel et al. 1996). Autophagocytosis is increased when cells have to face starvation (as reviewed by van den Hazel et al. 1996). Endocytosis is a transport mediated by vesicles, which is also important in the pathway of vacuolar proteolysis (van den Hazel et al. 1996). Three enzymes were described by (Achstetter, Wolf 1985) with location in the periplasmic space, between plasmalemma and cell wall. Two of these proteinases were metal dependent. Little was known about their function in the cell. At least 17 additional soluble proteinases were described by the two authors with unknown location. The soluble proteinases could probably be partially active outside the cell in the course of autolysis. This was proven for proteinase A in wine-like conditions (as reviewed by Charpentier and Feuillat 1993) A variety of membrane bound proteinases with unknown location were also described (Achstetter and Wolf 1985).

It could be concluded that Saccharomyces cerevisiae had a broad array of proteinases with different locations, activities, optima conditions and specificities. Achstetter and Wolf described the presence of at least 40 enzymes with proteolytic activity in Saccharomyces cerevisiae already in 1985. Protein degradation increased when yeast was exposed to stress such as nutrient limitations, heat and extreme pH values and vacuolar proteinase activity was increased when cells were exposed to mutagenic radiation (as reviewed by Achstetter and Wolf 1985; Hilt and Wolf 1992; van den Hazel et al. 1996). Levels of proteolytic activity depend also on the growth phase of the cells and level is higher when cultures were in stationary phase (van den Hazel et al. 1996). Proteolysis plays also a crucial role in sporulation of yeast and degradation of proteins damaged by heat, radiation or oxidative stress (Achstetter and Wolf 1985; Hilt and Wolf 1992). A part of the vacuolar proteases is in glycosylated state in form of precursor or as mature enzyme, e.g. proteinase A, B and Carboxypeptidase Y having molecular masses of 42, about 33 and 61 kDa respectively in the mature form (van den Hazel et al. 1996).

Besides unspecific proteolysis in the vacuole yeast seem to have also a system to degrade specifically proteins in the cytoplasm or the nucleus (as reviewed by Hilt and Wolf 1992; Sorokin et al. 2009). This proteolysis happens in the proteasome a complex multi protein structure always composed of a core unit of about 700 kDa and assocoiated with often two regulatory particles, which are also multi protein structures (Sorokin et al. 2009). Ubiquitin a small signal protein of 76 amino acids binds to target proteins and serves as a signal to proteolysis in the proteasome, but ubiquitin is removed before

(15)

14

proteolysis of the target molecule takes place. Proteolytic processing such as enzyme activation can also take place in the proteasome (as summarized by Sorokin et al. 2009). The expression of ubiquitin and of enzymes catalyzing binding of ubiquitin to target proteins is induced when yeast are confronted to starvation or heat. The proteasome has multiple (chymotrypsin-like, trypsin-like and peptidyl-glutamyl-peptide hydrolyzing) proteolytic activities. Caspase-like activity and threonine protease catalytic activity were also detected in the core unit of the proteasome (Sorokin et al. 2009). Proteolysis in the proteasome is possible in an ATP (adenosine triphosphate)-dependent (energy consuming) form but also in ATP-independent manner and in ubiquitin independent manner. Degradation of proteins in the proteasome seems to be important in case of short lived proteins and proteins that have become detrimental for the cell (van den Hazel et al. 1996).

1.3.1.3 Morphological changes in yeast cells during autolysis

Yeast cells in autolysis can also show morphological changes. The yeast cell wall as a whole persists during yeast autolysis (as reviewed by Babayan and Bezrukov 1985). Yeast cells performing the second fermentation of Champagne were observed by electron microscopy. The cells showed plasmolysis after three month and the inner layer of the cell wall composed mainly of ß-glucans disappeared, but cell walls persisted during fifteen years (Troton et al. 1989). Saccharomyces cerevisiae yeast showed cells that have shrunk after 5 hours of induced autolysis at 46°C in a light microscope, lost ovoid aspect and big vacuoles disappeared (Takeo et al. 1989). The cell wall formed granules after 5 hours of autolysis as observed under electron microscope (Takeo et al. 1989). Charpentier et al. (1986) observed by electron microscopy the cell wall of Saccharomyces species, which became wrinkled during induced autolysis in a wine-like medium. The work of Martinez-Rodriguez et al. (2001b) confirmed an ultra-structural change of the cell wall of Saccharomyces cerevisiae, which became rougher and got wrinkles during autolysis of yeast cells in wine-like medium or in sparkling wine. The volume and the extent of cytoplasm of the cell was reduced and cells showed a lot of small vesicles, which could be autophagosomes especially after 12 months of maturation of the yeast in a sparkling wine as observed under light microscope (Martinez-Rodriguez et al. 2001b).

It can be concluded that the cell wall of Saccharomyces yeasts is not completely broken down during autolysis, but it is modified. The cell wall of Saccharomyces cerevisiae has been extensively reviewed by Klis et al. (2002). It has to be kept in mind that the growth phase of a yeast culture influences cell wall composition. The cell wall is denser and thicker when cells are in stationary growth phase and stress factors like starvation, extreme temperatures or extreme pH seem also to induce strengthening of the cell wall structure e.g. the concentration in chitin increases. The inner part of the cell wall lying next to the cytoplasm membrane, consisting of ß-glucans, seemed to be mainly hydrolyzed during autolysis, at least, when autolysis took place in wine-like conditions at low temperature for months or years or at elevated temperatures for days. The glycosylated, mainly mannosylated proteins lying on the glucan layers seem to be less degraded (as reviewed by Charpentier and Feuillat 1993), but

(16)

15

proteolysis seemed also to occur in the cell wall when yeasts were autolysed in a wine-like medium at moderate (30°C) temperature (Charpentier et al. 1986).

1.3.1.4 Detection of yeast cells being in autolysis

Yeast colonies containing cells which are in autolysis can be detected by their release of alkaline phosphatase (Cabib and Duran 1975). Alkaline phosphatases were described as intracellular enzymes that were not released outside the cell and could be located bound on membranes or be in soluble form (Attias et al. 1970; Bauer and Sigarlakie 1975; Mitchell et al. 1981; Tonino and Steyn-Parvé 1963). Different researchers have used 5-bromo-4-chloro-3-indolylphosphate (BCIP) as substrate of alkaline phosphatase (Giovani and Rosi 2007; Gonzalez et al. 2003; Molero et al. 1993). BCIP can be integrated in complex solid yeast media and will detect yeast colonies having cells in autolysis by staining these colonies blue to turquoise.

1.3.1.5 Compounds released during autolysis

The autolysis of yeast cells leads to the release of intracellular compounds, mainly macromolecules and their degradation products into the surrounding medium (cf. review of Babayan and Bezrukov 1985). Proteins, peptides and amino acids were released during autolysis and the intracellular break down and liberation seemed to be the highest at temperatures between 40 and 50°C and at a slightly acidic pH (Hernawan and Fleet 1995; Kollar et al. 1993; Vosti and Joslyn 1954).

Glutamic acid, phenylalanine, leucine, alanine and arginine were the prevalent free aminoacids found in the autolysates (Hernawan and Fleet 1995). Alanine was also among the prevalent amino acids released during autolysis in wine-like conditions at moderate temperature (30°C) and also asparagine was liberated in higher amounts (Martinez-Rodriguez et al. 2001b; Perrot et al. 2002).

Most of the nitrogen compounds released during autolysis at moderate temperature (30°C) in a wine-like medium were peptides (Martinez-Rodriguez and Polo 2000; Martinez-Rodriguez et al. 2001a; Perrot et al. 2002). Perrot et al. (2002) found that amino acids and peptides were mainly released in nearly equal concentrations. Proteins were also liberated but in a more than hundred fold lower concentration. The peptides and proteins were mainly composed of glycine, glutamine and glutamic acid, alanine, lysine, asparagine and aspartic acid, proline, threonine and serine. Guilloux-Benatier and Chassagne (2003) confirmed that proteins made up only below 10% of the nitrogen compounds released during autolysis in wine-like media. They found mainly alanine, leucine, γ-aminobutyric acid and valine as amino acids liberated in free form and the basically the same amino acid composition in peptides and proteins as observed by (Perrot et al. 2002).

Degradation products of cellular DNA (deoxyribonucleic acid) and RNA (ribonucleic acid) have been found in autolysates of yeasts of the genus Saccharomyces (Hernawan and Fleet 1995; Kollar et al. 1993; Vosti and Joslyn 1954; Zhao and Fleet 2003, 2005). Products out of RNA predominated as the cells contained between 0.2 g/100 g and 1.5 g/100 g of DNA, but roughly ten times more RNA (Zhao and Fleet 2003, 2005). The conditions of autolysis determined which concentrations of the nucleotides

(17)

16

of 2’, 3’ or 5’ type were in the autolysate. 2’ nucleotides were not naturally present in the cell but formed by chemical modification. Nucleotides were the prevalent form of RNA degradation products. 5’IMP (inosine monophosphate) formed out of the precursor 5’AMP (adenosine monophosphate) and 5’GMP (guanosine monophosphate) were reported to be potent flavour enhancers as cited by (Zhao and Fleet 2003, 2005).

Polysaccharides and reducing sugars were also found in yeast autolysates of Saccharomyces yeasts (Hernawan and Fleet 1995; Kollar et al. 1993). Polysaccharides and glycosylated proteins may mainly derive from the partial degradation of the yeast cell wall cf. above.

Lipids were analyzed in yeast autolysates by Hernawan and (Fleet 1995). They found no phospholipids in the autolysates. Comuzzo et al. (2006) found 8.5 g/100 g of lipids in a yeast extract used in wine. Release of lipids was also investigated during autolysis in wine-like conditions. 2 to 9 % of the lipid material was set free during autolysis and triacyl glycerides and sterol esters prevailed (Pueyo et al. 2000).

1.3.1.6 Influence of yeast autolysis in wine on its chemical composition and sensory character Several reviews on yeast autolysis during wine and sparkling wine production have been published in the last 20 years (Alexandre and Guilloux-Benatier 2006; Charpentier and Feuillat 1993; Fornairon-Bonnefond et al. 2002).

In the following, only some effects of compounds released during yeast autolysis on wine’s character will be outlined according to the reviews mentioned and some other studies are also cited.

Peptides out of the proteolysis of a heat shock protein (Hsp P 12) were isolated from yeast autolysate and they increased the perception of sweetness in dry red wine (Marchal et al., 2011).

Amino acids released during autolysis can be precursors of aroma compounds found in some types of wine like threonine, which can be transformed to 3-hydroxy-4,5-dimethyl-2(5H)-furanone (as reviewed by Charpentier and Feuillat, 1993) also named sotolon. Sotolon seems to be especially a contributor to aroma in wines aged with flor yeast like “Sherry fino” and “vin jaune”(Martin et al. 1992). Mannoproteins that can be released from the yeast cell wall during autolysis (cf. above) can modulate the perception of bitterness caused probably by polyphenols in conditions similar to wine (Vidal et al. 2004). Furthermore mannoproteins can modify the volatility of aroma compounds found in wine, like norisoprenoids, alcohols and esters in wine-like conditions (Chalier et al. 2007; Lubbers et al. 1994b). Nucleotides, that are released from autolysing yeast cells, are well known as flavour enhancing compounds in the food industry and showed such an effect in champagne but only at concentrations higher than naturally present (Charpentier et al. 2005). The flavour enhancing effect of nucleotides can be reinforced by glutamic acid also set free during yeast autolysis (cf. above) and such a combined effect could not be ruled out in sparkling wines (Charpentier et al. 2005).

Lipids released during autolysis (cf. above) can set free fatty acids, which can be degraded to hexanoic, octanoic and decanoic acid. These fatty acids have aroma thresholds in the range of mg/l (as reviewed by Francis and Newton 2005) and are described as rather unpleasant (sweat, cheese).

(18)

17

On the other hand esters can be formed out of these fatty acids with alcohols naturally present in wine and these esters having fruity aroma notes can positively modulate wine’s aroma profile (Francis and Newton 2005). Higher alcohols and aldehydes were also released from yeast cells during autolysis (Chung as cited by Alexandre and Guilloux-Benatier 2006).

Furthermore the insoluble yeast lees that are in contact with wine during autolysis can adsorb various compounds and thus influence wine’s sensory profile. An example is the adsorption of polyphenolic compounds such as anthocyanins, phenolic acids but also proanthocyanidins (Mazauric and Salmon 2005, 2006; Vasserot et al. 1997). Their potential sensory influence of phenolic compounds will be outlined later in the section 1.3.4. Another important role of yeast lees is the protection of wine against oxidation by binding oxygen (Fornairon et al. 1999; Fornairon-Bonnefond and Salmon 2003) and by release of reductive compounds containing cysteine (Tirelli et al. 2010).

Yeast cell walls can adsorb typical wine aroma compounds such as methyl-2-propanol or ethyl esters of fatty acids. Yeast lees are also able to bind aroma compounds causing off flavours such as some thiols (Lavigne 1996; Vasserot et al. 2003) or volatile phenols (Pradelles et al. 2008).

1.3.2 Creation of yeast mutants showing inducible autolysis and their use during wine making Gonzalez et al. (2003) and Giovani and Rosi (2007) created mutants by UV (ultra violet) radiation out of Saccharomyces yeast strains, but Giovani and Rosi used spores in order to get a diploid strain in homozygous state. They screened the mutants for release of alkaline phosphatase, an autolysis marker (on BCIP medium cf. above), during a heat shock of 37°C. Mutants which were positive on BCIP-test showed abnormal cell morphology, like plasmolysis and cytoplasmic granules bigger than in the mother strain, after five days of heat shock (Gonzalez et al. 2003). Plasmolysis, deformation of cell walls and burst cells were also observed by Giovani and Rosi (2007) after heat shock of 37°C. Some of the mutants showed a higher release of proteins at temperatures equal to or higher than 25°C (until 37°C) and a higher liberation of amino acids at 25°C (Gonzalez et al. 2003). Nunez et al. (2006) worked with mutants of Gonzalez et al. (2003) and found a mutant liberating more polysaccharides with mannose as the main sugar compound during second fermentation of sparkling wine for nine month. The sparkling wine elaborated with this mutant had better foaming characteristics while not showing differences in the other sensory descriptors. Giovani and Rosi (2007) reported that some mutants released more polysaccharides during fermentation of a medium simulating grape juice than the mother strain at 28°C while preserving the same viability as the mother strain (Giovani and Rosi 2007).

1.3.3 Fining of must and wine

1.3.3.1 General considerations

Fining of must and especially wine is an old practice in wine making, which had traditionally the aim of clarification of wine and also of improvement of its sensory characters, like colour but also taste and astringency.

(19)

18

The process of fining consists of the addition of proteins, traditionally of animal origin, like egg white, casein of milk, isinglass or gelatine, to the wine or must. The proteins will be precipitated in the wine/must by tannins, which are naturally present in the wine or which are added in some cases, mainly in white and rosé wines. This precipitation is connected with a temporary increase in wine’s turbidity and with flocculation, often formation of flakes visible by eye. The precipitate of proteins and tannins will settle down in a later stage and other particles causing turbidity in the wine/must will connect to the precipitating protein-tannin composite. The turbidity of the liquid will be thus decreased. The speed of settlement by gravity of the tannin-protein-particles is dependent on several factors, namely the diameter of the particles, the difference of density between the liquid and the particle and also the viscosity of the liquid (in ideal conditions the law of Stokes would be valid; for further explanations the reader is referred to Ribéreau-Gayon et al. 2004b pp. 383-390).

The factor viscosity of the must/wine can limit a successful fining, i.e. complete settlement of particles causing turbidity. Must or wine can present high levels of viscosity caused by pectin (composed of polygalacturonic acid, polysaccharides and also glycoproteins) extracted out of the grapes or by glucans present especially in musts/wines that have been produced out of grapes infected with grey rot (Botrytis cinerea). Polysaccharides and glycoproteins can also act as “protecting colloids” avoiding interactions between proteins and tannins and thus inhibiting the formation of composites that are able to settle down by gravity (cf. Ribéreau-Gayon et al. 2004b pp. 383-390 and de Freitas et al. 2003). Tannins can be cause of excessive astringency and bitterness of wines and of colour deviations and their precipitation by the protein fining agents can thus in some cases improve wine’s sensory quality. This will be outlined in the following passages.

1.3.3.2 Composition of polyphenols of must and wine and their quantification or estimation The phenolic compounds in must and wines have been extensively reviewed in Ribéreau-Gayon et al. (2004b, pp.179-259), but the different categories of polyphenolic compounds will be mentioned and detailed in the following passage. The text of this passage is based on Ribéreau-Gayon et al. (2004b, pp.179-259), unless otherwise stated.

Phenolic acids and their derivatives are found in white must, white wines and red wines. They derive from benzoic acid or cinnamic acid and are found in concentrations of 10 to 30 mg/l in white must or wine and in ten times higher concentrations in red wines. Derivatives of the two phenolic acids are often glycosylated and derivatives of cinnamic acid are mostly esterified with tartaric acid.

A study of the concentrations of phenolic acids in German white wines of the varieties Riesling, Silvaner, Traminer and of Bacchus, Müller-Thurgau and Rieslaner, which are cross-breeds having Riesling as one parent, showed that all varieties had concentrations from 1 to 10 mg/l of the phenolic acids caffeic acid, ferulic acid and coumaric acid as well as their esters with tartaric acid (Pour Nikfardjam et al. 2007). Furthermore wines of all these varieties contained 10 to 20 mg/l tyrosol, which should be derived from yeast metabolism according to Ribéreau-Gayon et al. (2004b, p. 180).

(20)

19

Flavonoids are another important group of phenolic compounds in white must, wine and red wine and can be also found in the grape in glycosylated form. Red wine contains some 100 mg/l of flavonoids in form of aglycones and white wine around 100 times lower concentrations.

Anthocyanins are the red pigments in grapes and are only found in the skin of the grape berries except in case of teinturier varieties in which anthocyanins are also found in the pulp. They are present in glycosylated forms that are in most grape varieties also acylated with p-coumaric, caffeic or acetic acid. Anthocyanins have a positive charge in their heterocycle, which makes them chemically more unstable than flavonoids and their coloration is dependent on the surrounding medium, as far as pH and concentration of SO2 (sulphur dioxide)are concerned. Anthocyanins have red colour at acidic pH and decolorize when pH rises with a minimum colour around pH 3.2 to 3.5. Anthocyanins have a blue colour at pH values above 4 and at neutral or alkaline pH values their colour is yellow (for further explanations please refer to Ribéreau-Gayon et al. 2004b, pp.192-202). Furthermore anthocyanins are decolorized by sulphurous acid. Copigmentation, formation of anthocyanins-tannin-compounds by weak, non-covalent bonds, between anthocyanins and tannins occur in wine conditions. Condensation, binding reactions involving covalent stable bonds, between anthocyanins and condensed tannin of the grape (procyanidins) occur directly or via ethyl links. These anthocyanin-tannin-compounds can occur in colourless or coloured forms, having bluish, orange or yellow colour depending on their chemical structures. These condensed compounds of anthocyanins and tannins are not decolorized by sulphur dioxide and their concentrations increase during wine storage.

Tannins are by definition phenolic compounds that form stable composites with proteins. The word tannin is derived from “tanner” to tan, which means the preservation of animal hides by treating with tannins or mineral salts. Condensed tannins are found in the skin, the seeds and the stems of grapes and are polymers of 3-flavanols, also called catechins. Red wines are much richer in condensed tannins due to their localization in the grape berry and 1 to 4 g/l are found whereas white must and wines contain around ten times less tannins. (+) catechin and (-) epicatechin are the base units of condensed tannins, also called procyanidins, as these polymers decompose in hot acid media forming cyanidins of red colour. The catechin monomers are too small to form stable (mostly non-covalent) bonds with proteins, but proteins can form stable complexes with catechin dimers.

The structures of condensed tannins (procyanidins) are very complex. The monomers are connected via covalent bonds between the C4 and the C6 or C8- carbon atoms. Ethyl bonds formed by reaction of ethanal (acetaldehyde), out of coupled oxidation between ethanol of the wine and procyanidins, can also link catechin units within procyanidins. A similar type of bonds between catechin units have been also described with other aldehydes. Oligomeric procyanidins are composed of around two to ten monomer units and polymeric procyanidins of more than ten catechin units. Procyanidins are not found in glycosylated forms in wines. Procyanidins are subject of complex oxidation reactions and can scavenge free radicals. The polymers of these reactions have brown colour and precipitate in must or wine.

(21)

20

The degree of procyanidins can be assessed by liquid chromatography after mild acid hydrolysis of the catechin polymers and subsequent reaction with phenylmethanethiol (Rigaud et al. 1991) or phloroglucinol (Kennedy and Jones 2001). This allows distinguishing between terminal and intermediary catechin units and as a result a mean degree of polymerization can be calculated.

Polyphenols in must and wine have a lot of different chemical structures as can be seen in the former passage. The chemical characterization of polymeric procyanidin- and anthocyanin polymers still pose difficulties even with the modern chromatographic and other physico-chemical methods, such as NMR (nuclear magnetic resonance) and mass spectrometry.

Chemical indices have been proposed to estimate the concentrations of different classes of phenolic compounds and those which were used in the studies of this thesis are outlined below.

Absorbance at 280 nm is used to estimate concentrations of total polyphenols in red wines as flavonoid compounds have an absorbance maximum at that wavelength (Somers and Ziemelis 1985). The absorbance spectrum of white musts or wines in UV(ultraviolet light)-domain does not have a sharp absorbance maximum at 280 nm (Somers and Ziemelis 1985), but shows two broad absorbance maxima within the ranges 265-285 nm and 315-325 nm. A part of the absorbance at 280 nm in white musts and wine is caused by proteins naturally present (Somers and Ziemelis 1985). The second peak in UV-absorbance of white musts and wines is caused by non-flavonoid polyphenols, namely esters of derivatives of cinnamic acid with tartaric acid and also glucose. Somers and Ziemelis (1985) propose to characterize polyphenols of white musts and wines by the absorbance at 280 nm and at 320 nm but corrections are made for non-phenolic compounds. They suggest two options and the preferred option would be treating the must or wine sample with a massive dose of PVP (polyvinyl pyrrolidone) of 100 g/l, adsorbing all phenolic compounds, and measure it as a blank against the must or wine. The other option would be subtracting the fixed values determined in their studies, namely 1.4 from the absorbance at 320 nm and 4 of the absorbance at 280 nm. It has to kept in mind that the corrected absorbance measured at 280 nm is not only due to flavonoid compounds in white must or wine but that a part of this absorbance is due to the derivative of cinnamic acids present, which can be estimated to contribute a value of 2/3 of the corrected absorbance at 320 nm to the absorbance at 280 nm.

Total polyphenol-index (absorbance at 280 nm at 10 mm path length) ranged from 27 to 100 (mean of 54) in Australian red wines and the absorbance in white wines from 5 to 15 (mean of 8) (Somers and Ziemelis 1985). Absorbance values at 280 nm of 40 to 70 or 50 to 95 in red wines were found by Vivas et al. (2003) and Iturmendi et al. (2010).

The astringency of tannins in red wine can be estimated by their reaction with “model proteins” and bovine serum albumin (BSA) was used to assess the “tannic power” of tannins of red wine (de Freitas 1995, de Freitas et al. 2003). De Freitas et al (de Freitas,1995; de Freitas et al. 2003) used nephelometry to follow the formation of tannin-BSA-precipitates in model solutions similar to wine. Kennedy et al. (2006) stated that the absorbance at 280 nm, an index of the concentration of total polyphenols in red wine did not allow statements on wine’s astringency. A “BSA-index” following the

(22)

21

principles of de Freitas (1995) and de Freitas et al. (2003) could however predict astringency of red wines (Kennedy et al. 2006).

The colour intensity of red wines can be estimated by measuring absorbance at 420 nm (absorbance of yellow colour), 520 nm (absorbance of red colour) and 620 nm (absorbance of blue colour) according to Glories (1984). The colour intensity would then be the sum of the absorbance at the three wavelengths. The quotient of absorbance 420 nm / absorbance 520 nm gives an estimation of the hue of the wine. Red wines having a high hue have a more orange to brownish colour and wines with a low hue show a redder colour, but the total colour intensity has also to be considered. Colour intensity values of red wines cover a wide range from 5 found in Pinot noir (Charpentier et al. 2006) to 12 to 14 in Bordeaux wines (Charpentier et al. 2006; Glories 1984) or Spanish red wines (Iturmendi et al. 2010) and 30-50 in some Portuguese wines (Castillo-Sanchez et al. 2008; Cosme et al. 2009).

The concentration of anthocyanins in red wine can be estimated by measuring their red colour (absorbance at 520 nm) in acidic medium and to compare this with the same acidified sample that is decolorized with sulphite. It has to be kept in mind that free anthocyanins and a part of the anthocyanins-tannin compounds are quantified by this method (Glories 1984). PVPP index estimates the proportion of anthocyanins combined with tannins of the whole concentration of anthocyanins measured by decolorization with sulphite (Glories 1984). This method is based on the adsorption of all polphenolic compounds on polyvinyl polypyrrolidone (PVPP) and the subsequent elution of free anthocyanins with an acidified alcoholic solvent. The concentration of free anthocyanins eluted from PVPP and dissolved in model wine is then quantified by decolorization with sulphite (cf. above). The yellow colour of white must and wine can be estimated by measuring the absorbance at 420 nm, but white must or wine does not have a clear absorbance maximum in light of wavelengths visible by the human eye (Ribéreau-Gayon et al. 2004b pp. 253-255). White wines showed absorbances at 420 nm from 0.05 to 0.1 in case of Pinot blanc wine (Vrhovsek and Wendelin 1998) to 0.3 in a Portuguese white wine (Cosme et al. 2008).

The sensory effect of phenolic compounds on wine, mainly red wine, has been studied by several groups. First, the works of de Freitas (1995), de Freitas et al. (2003) and Kennedy and Jones (2001) using the precipitation reaction between BSA (bovine serum albumin) and polyphenols for modeling astringency have to be mentioned, which have been already detailed. Furthermore phenolic compounds were separated by chromatographic methods and the fractions were evaluated by sensory analysis in the works of Hufnagel and Hofmann (2008), Sun et al. (2013), Vidal et al. (2004) and Weber et al. (2013).

Polymeric proanthocyanidins were described as causing astringency (Hufnagel and Hofmann 2008; Sun et al. 2013; Vidal et al. 2004), whereas Weber et al. (2013) described oligomeric proanthocyanidins as having a higher astringency than polymeric proanthocyanindins and a fraction containing anthocyanins and other phenolic compounds. Hufnagel and Hofmann (2008) described flavan-3-ols and ethylesters of hydroxybenzoic or hydroxycinnamic acids as contributing to bitterness.

(23)

22

Other derivatives of lower molecular weight of hydroxybenzoic and hydroxycinnamic acid were described as astringent as well as glycosides of falvanol and flavonal (Hufnagel and Hofmann 2008). Vidal et al. (2004) pointed out that there are interactions between the alcohol concentration and the perception of bitterness of phenolic compounds and that glycoproteins can diminish the perception of bitterness or astringency of polyphenolic compounds (Vidal et al. 2004).

1.3.3.3 Interactions between proteins and polyphenols

The interaction between proteins and tannins is a key feature in the process of fining and will be outlined in the following. A review on theories of fining wine with proteins can be found in Ribéreau-Gayon et al. (2004b pp. 390-398) and some points of this review will be detailed in the following. The first studies interpreted the clarification of wine by fining as aggregation of particles having opposite charges. Particles in the wine causing turbidity had a negative charge and gelatine molecules a positive charge as detected by electrophoresis. Their aggregation and at least partial neutralization of charges should be the mechanism underlying clarification (Rüdiger and Mayr 1928 and 1929 as cited by Ribéreau-Gayon et al. 2004b, p. 390).The studies of Ribéreau-Gayon of the 1930’s (summarized in Ribéreau-Gayon et al. 2004b, pp. 391 and 392) put up another type of mechanism. Flocculation between fining proteins and tannins of red wine should happen in two stages in that model. First, protein molecules having positive charges in wine conditions should be surrounded by tannin molecules carrying negative charges. That should result in still soluble complexes that are in the second stage discharged by metal cations leading to flocculation and sedimentation.

Modern theories of fining focus more on chemical interactions between wine colloids and proteins of the fining agent. Lagune (1994, as cited in Ribéreau-Gayon et al. 2004b, pp. 392-398) studied the fining of red wine with gelatine. Proteins and tannins carry electric charges on the surface of the molecules under specific conditions (pH, temperature etc.). The sum of positive and negative charges in a solution with charged particles like proteins or tannins or red wine can be characterized by its initial streaming potential and by the charge density of the system (refer to Ribéreau-Gayon et al. 2004b, pp. 392-398 for further explanations). The streaming potential is the potential created by a moving piston in the solution. The charge density is the molar concentration of a polyelectrolyte of opposite charge compared to the initial streaming potential that is needed to completely neutralize this potential. The initial streaming potential of red wines is negative, but not proportional to its tannin concentration as other particles in the red wine like polysaccharides may also carry negative charges. The volume of a gelatine solution (having a positive initial streaming potential) necessary to neutralize the initial streaming potential of a red wine was a lot higher than the gelatine concentrations used for a successful fining/clarification of the corresponding wine. Fining does consequently not rely on the complete neutralization of charges in the wine.

Interactions between proteins and tannins have been extensively studied in the last 40 years and a review is given by Ribéreau-Gayon et al. (2004b, pp.398-402). The nature of binding between proteins and tannins seems to rely mainly on non-covalent bonds such as hydrogen bonding and

(24)

23

hydrophobic binding, but also covalent binding has been reported (as summarized by Ribéreau-Gayon et al. 2004b, pp. 398-401 and Ricardo da Silva et al. 1991). The chemical conformation and concentration of tannin and protein molecules, the proline concentration in the protein molecules and the hydrophobicity of the tannin molecules seem to influence tannin-protein-interactions (as summarized by Ribéreau-Gayon et al. 2004b, pp. 398-401 and Ricardo da Silva et al. 1991).

The glycosylation of proteins can also influence the protein-tannin-interactions (reviewed in Ricardo da Silva et al. 1991). Furthermore factors of the environment of the tannin and protein molecules such as pH, temperature, ionic strength and ethanol concentration seem to influence their interaction (as summarized by Ribéreau-Gayon et al. 2004b, pp. 398-401 and Ricardo da Silva et al. 1991).

Some important studies mainly of the last twenty years will be outlined in the following.

Procyanindin trimers bound more poly-L-proline than dimers and most binding was observed for procyanindin trimers that showed esterification with gallic acid (Ricardo da Silva et al. 1991).

In the latter cases equivalent masses of poly-L-proline and tannin were precipitated and all tannins were precipitated (Ricardo da Silva et al. 1991).The length of proline chains also influenced their binding capacity showing a higher binding efficiency above 19 kDa. The imino acid proline shall preferably form hydrogen bonds with hydroxy rests of the phenolic compounds as it contains an oxygen atom adjacent to secondary amine nitrogen (Ricardo da Silva et al. 1991).

Binding reactions (chemical interactions resulting in release of energy) have been observed for galloylated procyanidin monomers, but not for monomeric catechins, with poly-L-proline (Poncet-Legrand et al. 2007b). This release of energy should be caused by formation of hydrogen bonds, but hydrophobic interactions seemed to be also involved as galloylation increased the hydrophobic character of the procyanidin (Poncet-Legrand et al. 2007b). Gelatine and casein bound also preferably with galloylated procyanidin trimers, but gelatine bound at maximum 30 % of procyanidins and casein 50 % of procyanidins when used in a 3.7 higher concentration than gelatine (Ricardo da Silva et al. 1991). A higher number of o-dihydroxygroups in the molecule of procyanidins shall increase their binding capacity. The lower procyanidin concentrations bound by gelatines and casein compared to L-poly-proline should be partly due to their lower concentrations of proline of at maximum 25 g/100 g in case of gelatine or 12 g/100 g in case of casein (Ricardo da Silva et al. 1991). The proline concentration shall also be an important factor in formation of protein-polyphenol-haze in beverages, as well as the degree of polymerization of procyanidins (as reviewed by Siebert 1999). Hydrophobic interactions and hydrogen bonding seem to be the forces combining haze-active polyphenol-protein aggregates and an increase in pH from 3 to 4 favoured haze formation (as reviewed by Siebert 1999). An extensive study on binding of procyanidins with gelatine and their hydrolysates did not show an influence of oxygenation or ionic strength on tannin binding (Yokotsuka and Singleton 1987).

Procyanidin dimers and oligomers bound a higher concentration of gelatine molecules of different molecular weights in the range from 70 to 2 kDa when pH value increased from pH 3 to 4 (Yokotsuka, Singleton 1987, 1995) which was also stated by Siebert (1999).

(25)

24

A decrease of temperature during incubation of procyanidin - gelatine mixtures of different molecular masses from 70 Da to 2 kDa from 25 to 15 °C increased procyanidin precipitation (Yokotsuka and Singleton 1987). This behavior should be caused by prevalent hydrogen bonding connecting procyanidins with gelatine molecules.

The molecular weight of gelatine fractions had no influence on their binding capacity with procyanidins. 80 % of the tannins were bound at maximum at equal concentrations of gelatine and tannin (Yokotsuka and Singleton 1987). Precipitation experiments with BSA (bovine serum albumin) and tannins of grape seeds also showed that a complete depletion of the tannins in the mixtures was not possible (Schmauch 2010). Gelatine fractions having molecular masses of 10 to 2 kDa readily precipitated with procyanidin dimers or oligomers (Yokotsuka and Singleton 1987). A complete precipitation of the proteins/peptides was only achieved when gelatine of an average molecular mass of 70 kDa was added to polymeric or oligomeric procyanidins in a ratio of 1:1.

Catechin monomers were not able to precipitate with the gelatine fractions in the study of (Yokotsuka and Singleton 1987), which confirms the statement of Ribéreau-Gayon et al. (2004b).

Glycosylation can also influence the binding capacity of proteins with tannins. An arabinogalactan-protein, a type of protein found in pectin of grapes, showing a low protein concentration of 6 g/100 g and 90 g/100 g of neutral sugars bound procyanidins comparably to gelatines of molecular weights around 10 kDa in the study of Ricardo da Silva et al. (1991). The mannoprotein invertase showed a 10 times lower binding capacity for procyanidins than the non-glycosylated protein BSA (bovine serum albumin) (Rowe et al. 2010).The aggregation of PRP (proline-rich proteins) of human saliva was reduced by their glycosylation and aggregation started at higher proportional ratio of procyanidins to glycosylated PRP. The aggregates were also of smaller size in case of glycosylated PRP (Pascal et al. 2008; Sarni-Manchado et al. 2008). The aggregation of procyanidin molecules can be inhibited by the presence of mannoproteins from yeast (Charpentier et al. 2004; Poncet-Legrand et al. 2007a; Riou et al. 2002). Polysaccharides can inhibit the precipitation of oligomeric procyanidins with the protein BSA and ionic polysaccharides such as xanthan and pectin were more efficient (de Freitas et al. 2003). Mannoproteins of yeast can also inhibit the formation of haze by wine proteins (Dupin et al. 2000 b; Dupin et al. 2000 a; Waters et al. 1994). This haze formation in wine was attributed to protein-polyphenol interactions by Siebert (1999) as outlined above. On the other hand glycoproteins containing mannose could precipitate tannins of red wine (Gambuti et al. 2012) or possibly precipitate procyanidins and colour pigments of red wine (Guadalupe and Ayestaran 2008).

Negative charge densities in a range of 30 to 120 milli equivalents (meq)/g have been found in condensed tannins consisting of polymerized procyanidins (Lagune 1994 as cited by Ribéreau-Gayon et al. 2004bp. 400), whereas very low, negligible charges have been found on tannins of grape seed at pH of wine (3.5) by Vernhet et al. (1996). The charge density of a protein fining agent, its amino acid composition and its molecular masses determine also its fining capacities (as summarized in Ribéreau-Gayon et al. 2004b, p. 400).

(26)

25

No molar ratio between the concentration of protein fining agent (i.e. gelatine) and the concentration of eliminated tannin was found (Lagune, 1994 as cited by Ribéreau-Gayon et al. 2004b p. 400). The pH value of the wine influenced the speed of flocculation of fining proteins and more tannin was eliminated at higher pH values. The presence of metal cations was described as indispensable of flocculation between tannins and proteins (all as reviewed by Ribéreau-Gayon et al. 2004b, pp. 400 and 401).

The perception of astringency seems to be related to interactions between human saliva and polyphenols of red wine (de Freitas and Nunez 2001; Schmauch 2010). Saliva has proteins rich in proline called PRP and these PRP are precipitated by procyanidins. The precipitation is already possible with monomers. The intensity of formation of precipitates analyzed as increase in turbidity of procyanidins-PRP-mixtures depended on the stereochemistry of the procyanidins, like position of hydroxyl groups on phenolic rings, type of monomer bonds and galloylation (esterification with gallic acid) (de Freitas and Nunez 2001). The conformation of proteins also influenced the intensity of precipitation. PRP having randomly coiled structures produced more intense turbidity at lower concentrations than α-amylase having a globular structure (de Freitas and Nunez 2001). Bovine serum albumin (BSA) did not bind procyanidin dimers and trimers at pH 5 (de Freitas and Nunez 2001), but did bind to commercial tannins of grape seed. The maximum amount of binding and precipitation was observed at pH 3.5 by nephelometry when BSA was in F-form, in which a part of the interdomain helices are disrupted catalyzed by acid pH (Nakamura et al. 1997). High binding activity between BSA and tannins was still observed at pH 5, a pH near pH of mucus in oral cavity (de Freitas, Nunez 2001) at which BSA is at its isoelectric point (Schmauch 2010). Different red wines precipitated all proteins found in human saliva and Schmauch (2010) raised the question if the perception of astringency was only related to complete precipitation of saliva proteins resulting in a lack of lubrication of the oral cavity. He put up the hypothesis that direct interactions between proteins of mucous membranes of the oral cavity may also be a cause of the perception of astringency (Schmauch 2010). The presence of polysaccharides, mainly ionic ones like xanthan and pectin, diminish reactivity of procyanidin oligomers with BSA at pH similar to saliva, which could also modulate perception of astringency (de Freitas et al. 2003). Mannoproteins seem also to reduce astringency of red wine (Escot et al. 2001) and can reduce bitterness probably caused by polyphenols in mixture with arabinogalactan proteins in wine-like conditions (Vidal et al. 2004).

1.3.3.4 Selected studies on fining of must and wine and characterization of protein fining agents

The studies shown in this passage used mainly the protein fining agents gelatine, casein, isinglass and yeast protein extracts (YPE) also examined in the research of this thesis.

The study of Ricardo da Silva et al. (1991) showed that fining red wine of the variety Mourvèdre with the same concentrations of gelatine or casein used in the binding study with procyanidins resulted in no decrease in concentrations of procyanidin dimers and trimers. The same fining agents bound

(27)

26

however these phenolic compounds in synthetic wine solution (cf. above), which may be due to the fact that more reactive polymeric procyanidins or anthocyanins compounds “protected” the procyanidin di- and trimers. It can be concluded that interaction studies between procyanidins and proteins in wine-like conditions are useful for the study of chemical interactions but cannot be a model for the reactions in wine in any case. The colour intensity of this Mourvèdre wine diminished slightly by around 5 % as well as the index of total polyphenols (absorbance at 280 nm) (Ricardo da Silva et al. 1991).Gelatine of mean molecular weights between 20 and 10 kDa reduced turbidity of Pinot noir red wine from 70 NTU (nephelometric turbidity units), a noticeable turbidity, to 30 NTU, a still slight turbidity, when used at 8 g/hL (Marchal et al. 2002a). The sensory character of the wine was however not changed by this potential reduction in tannin concentration (Marchal et al. 2002a).

Gelatines of different mean molecular masses ranging from 10 to 66 kDA bound preferably proanthocyanidins of red wine and their concentration was reduced at maximum by 9 % (Sarni-Manchado et al. 1999). Gelatine of a mean molecular mass of 25 kDa showed the highest effect and the mean degree of polymerisation of tannins precipitated by gelatine was higher than that of the wine and galloylated tannins were in particular precipitated (Sarni-Manchado et al. 1999).The authors pointed out that soluble protein-tannin complexes were formed, mainly when fining was done with gelatine of high molecular weight. A liquid gelatine preparation of a mean molecular mass of 25 kDa and two fractions of this gelatine of 16 and 190 kDa did not bind monomeric proanthocyanidins, but condensed tannins, i.e. proanthocyanidins to 10 to 15 %. The original gelatine had the highest effect (Maury et al. 2001) and precipitated mainly tannins of high molecular mass, as well as the gelatine fractions. The gelatine fraction of lower molecular mass of 16 kDa bound tannins of a higher mean degree of polymerization than the original gelatine and the high molecular mass fraction. Furthermore galloylated tannins were preferentially bound to gelatine molecules. The concentration of tannins of red wines after fining was lower after fining with original gelatine and its two fractions, but only two of four wines showed a lower mean degree of polymerization after fining with the gelatine fraction of low molecular mass. A part of the wines was tasted after fining with the gelatine fractions and astringency was decreased, which was related with a decrease in total proanthocyanidins in both wines. A decrease of the mean degree of polymerization reduced astringency only in one of two wines fined with the gelatine fraction of lower molecular weight of 16 kDa.

Gelatine at a concentration until 10 g/hl, casein until a concentration of 60 g/hl and isinglass at a concentration until 2 g/hl did not decrease the concentrations of total polyphenols, phenolic acids and catechins in white or red wine in the study of Fischerleitner et al. (2003).

Highly hydrolyzed gelatine with proteins of masses below 14 kDa reduced tannin fractions of mean degree of polymerization of 1.5, 3.4 and 4.9 of a red wine more or as only one when compared to a less hydrolyzed gelatine used at the same concentration (Cosme et al. 2009). This more hydrolyzed gelatine also reduced colour intensity of the red wine. The mean degree of polymerization of the

(28)

27

corresponding red wine was not reduced by fining with gelatine. Castillo-Sanchez et al. (2008) reported that fining red wines with 20 g/hl gelatine reduced their colour intensity.

First studies on yeast protein extracts (YPE) showed that YPE at concentrations between 20 and 50 g/hl clarified red wines to a similar extent than fining with liquid or solid gelatine preparations of concentrations of 6 g/hl or 30 to 100 ml/hl (Charpentier et al. 2006; Iturmendi et al. 2012).

The colour intensity of red wines of Burgundy or Bordeaux was reduced by YPE in a way comparable to fining with gelatine (Charpentier et al. 2006).

Patatin, a glycoprotein out of potato, was used for fining and did not reduce colour intensity of red wine at the same concentration as gelatine, which caused a loss of colour (Gambuti et al. 2012). On the other hand both fining agents reduced concentrations of total polyphenols and tannins, which was related with a decrease in astringency.

Fining with casein, isinglass or gelatine did not reduce the tendency of browning of white wine, a phenomenon that should be at least partly due to polyphenol oxidation, as stated in the studies of Fischerleitner et al. (2003) and Vrhovsek and Wendelin (1998).Vrhovsek and Wendelin (1998) also found that concentrations of esters of tartaric acid with cinnamic acids were not reduced by treatment with casein or gelatine. Some types of gelatine, isinglass and casein were however reported to decrease the concentrations of flavanol monomers and procyanidin oligomers and polymers in the study of Cosme et al. (2008). Furthermore the mean degree of polymerization of wine’s procyanidins was decreased after fining with casein, isinglass or gelatine. The turbidity of white must or wine could be reduced by fining with isinglass and casein as found by Cosme et al. (2008) and Marchal et al. (2002b). Caseins reduced also the yellow colour of white wine as stated by Cosme et al. (2008) and Vrhovsek and Wendelin (1998).

Only three studies were found that evaluated the sensory effect of fining (Gambuti et al. 2012; Marchal et al. 2002a; Maury et al. 2001 cf. above). Most of the articles cited in this section/passage stated that fining agents precipitated preferably oligomeric and polymeric proanthocyanidins. These proanthocyanidins were described as participating in the sensory impression of astringency as outlined by Gambuti et al .( 2012), Hufnagel and Hofmann (2008), Marchal et al. (2002a), Maury et al. (2001), Sun et al.(2013) and Vidal et al. (2004). Fining could thus influence wine’s astringency. It is also important to mention that wine colour and limpidity can be influenced by fining as shown in this section and these parameters are also part of the sensory perception of wines.

Protein fining agents should be characterized in a physico-chemical way to allow the winemakers to obtain effective fining agents and to avoid that their use could introduce material to wine that could be detrimental in any way to human health. Consequently, the International Organization of Vine and Wine (OIV) set up maximum concentrations of minerals, heavy metals and microorganisms in gelatine, casein and isinglass preparations, mostly to avoid the commercialization of products having negative influence on human health (cf. resolutions in International Oenological Codex of International Organization of Vine and Wine, 2012). On the other hand their protein concentration is